Will AI ever fully replace healthcare providers?

Tasha Community Manager, Dailynista admin
edited October 2022 in Discussion

@sventy recently spoke about the use of AI in medicine, which is why this article caught my eye.

I’m bummed the reporter arrived at the scenario where Alexa analyzes and diagnoses you by your voice. That sounds terrible, like how would Alexa deliver the news?  

The healthcare industry already gets criticism for a decline in bedside manor. Wouldn’t real progress mean AI is used as a tool by patients and healthcare providers to connect and get the care they need? 

I'm curious what others think. Will AI ever fully replace healthcare providers?


  • Lazer
    Lazer Dailynista
    edited October 2022

    I don't believe AI will ever fully replace healthcare providers, but the topic makes me think of an experience I had a few years ago with a "robot-ish"-assisted treatment session. Advance notice: this is not related to video, but to taking portions of treatment out of human hands.

    I'm very afraid of needles, to the point where I've blacked out and even gotten sick a few times in the past during treatment or testing that involved a needle. The dentist needle is the worst. So when a few years ago I was told I had to get a wisdom tooth extracted, I was terrified. In panicked Googling about what kind of needle would be used or if "laughing gas" was an option to help, I came across this thing called "The Wand".

    The Wand, I learned, is a device with a very tiny needle that delivers anesthetic to a very precise area, with a flow that is controlled by the machine itself and not by the doctor's hand on a plunger. It took a while to find a dentist who had it available, but in the end I found it at a place specifically tailored to people with needle or dental fear. 

    I'm sure the doctor controls the volumes of anesthetic and such and don't know the full details, but just knowing that the needle was so small and that the actual flow of the substance would be out of human hands made me feel better about the whole thing. I trusted a robot more than a human to control the injection. And it worked! Not only was it much less painful than I remembered (and my whole mouth didn't go numb), but I was also generally much less freaked out about the whole thing. Plus, unlike prior dentist injections, the flow seemed so smooth and even that I didn't even feel the anesthetic going into my gum.

    The procedure itself was very difficult - they had to remove a bit of my jaw bone and do a lot of yanking and drilling. The dentist said it was one of the tougher extractions she's done and the recovery was brutal. But as a patient, that procedure was a better experience than any cavity I remembered getting filled as a teenager just with the difference in the injection experience. I'm definitely only going to dentists who have this kind of machine-assisted injection from now on.

    Tl;dr: Sometimes robots do it better.

  • jayne
    jayne Member, Daily Alumni admin

    A bit related to that Alexa post, I recently read something about the new iPhone's car crash detection feature. In theory, your phone or watch should alert emergency services when it detects that you're in a severe accident. But it has alerted the emergency services a bit too much for people who are just on a roller coaster.

    Back to your question, I don't think AI will fully replace healthcare providers any time soon, but it will help a lot with diagnosis. I just hope that they don't base the AI training on the biased research that doctors often use.

  • I don't think we'll get there completely any time soon, in the sense that human providers would become obsolete. But I do think that diagnosis of certain conditions could become automated to a very large degree, with some human oversight. Likewise, as @Lazer points out, we're seeing certain medical procedures being augmented with technology that helps reduce errors and minimize the invasiveness of those procedures.

    What's really interesting to me is the degree to which privacy is impacted in developing software that can diagnose medical conditions: training machine learning systems requires a lot of data, and in this case, the data required is extremely sensitive. And @jayne made a great point above about the biases that are reflected in current medical research. It will be fascinating to watch how this area evolves in the coming years.